5c_job_failures_analysis.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Section 5c: Job Failure Analysis"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "from util.read_and_print_df import *\n",
    "from util.handle_nan import *\n",
    "from util.extract_json_attributes import *\n",
    "import matplotlib.pyplot as plt\n",
    "import matplotlib\n",
    "import pandas as pd\n",
    "from datetime import date, datetime, time, timedelta\n",
    "import time, datetime, pytz\n",
    "from datetime import datetime"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "plt.rcParams['text.usetex'] = True\n",
    "plt.rcParams['font.family'] = 'libertine'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>id</th>\n",
       "      <th>submit_date</th>\n",
       "      <th>start_date</th>\n",
       "      <th>end_date</th>\n",
       "      <th>node</th>\n",
       "      <th>nodetypes</th>\n",
       "      <th>numnodes</th>\n",
       "      <th>numcores</th>\n",
       "      <th>state</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>2021-12-26 17:11:18</td>\n",
       "      <td>2021-12-26 23:06:31</td>\n",
       "      <td>2021-12-31 23:06:50</td>\n",
       "      <td>r13n5</td>\n",
       "      <td>normal</td>\n",
       "      <td>1</td>\n",
       "      <td>16</td>\n",
       "      <td>TIMEOUT</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2</td>\n",
       "      <td>2021-12-26 17:11:18</td>\n",
       "      <td>2021-12-26 23:06:43</td>\n",
       "      <td>2021-12-31 23:06:50</td>\n",
       "      <td>r14n27</td>\n",
       "      <td>normal</td>\n",
       "      <td>1</td>\n",
       "      <td>16</td>\n",
       "      <td>TIMEOUT</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>3</td>\n",
       "      <td>2021-12-26 17:11:18</td>\n",
       "      <td>2021-12-26 23:06:43</td>\n",
       "      <td>2021-12-31 23:06:50</td>\n",
       "      <td>r15n12</td>\n",
       "      <td>normal</td>\n",
       "      <td>1</td>\n",
       "      <td>16</td>\n",
       "      <td>TIMEOUT</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>4</td>\n",
       "      <td>2021-12-26 17:11:18</td>\n",
       "      <td>2021-12-26 23:06:43</td>\n",
       "      <td>2021-12-31 23:06:50</td>\n",
       "      <td>r10n14</td>\n",
       "      <td>normal</td>\n",
       "      <td>1</td>\n",
       "      <td>16</td>\n",
       "      <td>TIMEOUT</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>5</td>\n",
       "      <td>2021-12-26 17:11:18</td>\n",
       "      <td>2021-12-26 23:06:43</td>\n",
       "      <td>2021-12-31 23:06:50</td>\n",
       "      <td>r10n30</td>\n",
       "      <td>normal</td>\n",
       "      <td>1</td>\n",
       "      <td>16</td>\n",
       "      <td>TIMEOUT</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1596958</th>\n",
       "      <td>2323559</td>\n",
       "      <td>2022-11-01 13:55:57</td>\n",
       "      <td>2022-11-01 13:55:57</td>\n",
       "      <td>2022-11-01 13:58:54</td>\n",
       "      <td>r28n1</td>\n",
       "      <td>gpu_titanrtx_shared</td>\n",
       "      <td>1</td>\n",
       "      <td>6</td>\n",
       "      <td>CANCELLED</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1596959</th>\n",
       "      <td>2323560</td>\n",
       "      <td>2022-11-01 13:56:12</td>\n",
       "      <td>2022-11-01 13:56:17</td>\n",
       "      <td>2022-11-01 13:56:24</td>\n",
       "      <td>r13n10</td>\n",
       "      <td>shared</td>\n",
       "      <td>1</td>\n",
       "      <td>6</td>\n",
       "      <td>FAILED</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1596960</th>\n",
       "      <td>2323561</td>\n",
       "      <td>2022-11-01 13:56:45</td>\n",
       "      <td>2022-11-01 13:56:45</td>\n",
       "      <td>2022-11-01 13:58:12</td>\n",
       "      <td>r28n1</td>\n",
       "      <td>gpu_titanrtx_shared</td>\n",
       "      <td>1</td>\n",
       "      <td>6</td>\n",
       "      <td>FAILED</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1596961</th>\n",
       "      <td>2323557</td>\n",
       "      <td>2022-11-01 13:54:21</td>\n",
       "      <td>2022-11-01 13:58:19</td>\n",
       "      <td>2022-11-01 13:59:01</td>\n",
       "      <td>r14n20</td>\n",
       "      <td>normal</td>\n",
       "      <td>1</td>\n",
       "      <td>16</td>\n",
       "      <td>COMPLETED</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1596962</th>\n",
       "      <td>2323558</td>\n",
       "      <td>2022-11-01 13:54:26</td>\n",
       "      <td>2022-11-01 13:58:19</td>\n",
       "      <td>2022-11-01 13:59:18</td>\n",
       "      <td>r14n21</td>\n",
       "      <td>normal</td>\n",
       "      <td>1</td>\n",
       "      <td>16</td>\n",
       "      <td>COMPLETED</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>1596963 rows × 9 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "              id         submit_date          start_date            end_date  \\\n",
       "0              1 2021-12-26 17:11:18 2021-12-26 23:06:31 2021-12-31 23:06:50   \n",
       "1              2 2021-12-26 17:11:18 2021-12-26 23:06:43 2021-12-31 23:06:50   \n",
       "2              3 2021-12-26 17:11:18 2021-12-26 23:06:43 2021-12-31 23:06:50   \n",
       "3              4 2021-12-26 17:11:18 2021-12-26 23:06:43 2021-12-31 23:06:50   \n",
       "4              5 2021-12-26 17:11:18 2021-12-26 23:06:43 2021-12-31 23:06:50   \n",
       "...          ...                 ...                 ...                 ...   \n",
       "1596958  2323559 2022-11-01 13:55:57 2022-11-01 13:55:57 2022-11-01 13:58:54   \n",
       "1596959  2323560 2022-11-01 13:56:12 2022-11-01 13:56:17 2022-11-01 13:56:24   \n",
       "1596960  2323561 2022-11-01 13:56:45 2022-11-01 13:56:45 2022-11-01 13:58:12   \n",
       "1596961  2323557 2022-11-01 13:54:21 2022-11-01 13:58:19 2022-11-01 13:59:01   \n",
       "1596962  2323558 2022-11-01 13:54:26 2022-11-01 13:58:19 2022-11-01 13:59:18   \n",
       "\n",
       "           node            nodetypes  numnodes  numcores      state  \n",
       "0         r13n5               normal         1        16    TIMEOUT  \n",
       "1        r14n27               normal         1        16    TIMEOUT  \n",
       "2        r15n12               normal         1        16    TIMEOUT  \n",
       "3        r10n14               normal         1        16    TIMEOUT  \n",
       "4        r10n30               normal         1        16    TIMEOUT  \n",
       "...         ...                  ...       ...       ...        ...  \n",
       "1596958   r28n1  gpu_titanrtx_shared         1         6  CANCELLED  \n",
       "1596959  r13n10               shared         1         6     FAILED  \n",
       "1596960   r28n1  gpu_titanrtx_shared         1         6     FAILED  \n",
       "1596961  r14n20               normal         1        16  COMPLETED  \n",
       "1596962  r14n21               normal         1        16  COMPLETED  \n",
       "\n",
       "[1596963 rows x 9 columns]"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df = pd.read_parquet(path_job_dataset)\n",
    "df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>id</th>\n",
       "      <th>submit_date</th>\n",
       "      <th>start_date</th>\n",
       "      <th>end_date</th>\n",
       "      <th>node</th>\n",
       "      <th>nodetypes</th>\n",
       "      <th>numnodes</th>\n",
       "      <th>numcores</th>\n",
       "      <th>state</th>\n",
       "      <th>split_nodes</th>\n",
       "      <th>...</th>\n",
       "      <th>submit_day_of_week</th>\n",
       "      <th>month_of_year</th>\n",
       "      <th>week_of_year</th>\n",
       "      <th>day_of_year</th>\n",
       "      <th>hour_of_year</th>\n",
       "      <th>minute_of_year</th>\n",
       "      <th>second_of_year</th>\n",
       "      <th>job_waiting_time</th>\n",
       "      <th>job_duration</th>\n",
       "      <th>rack</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>2021-12-26 17:11:18</td>\n",
       "      <td>2021-12-26 23:06:31</td>\n",
       "      <td>2021-12-31 23:06:50</td>\n",
       "      <td>r13n5</td>\n",
       "      <td>normal</td>\n",
       "      <td>1</td>\n",
       "      <td>16</td>\n",
       "      <td>TIMEOUT</td>\n",
       "      <td>{r13n5}</td>\n",
       "      <td>...</td>\n",
       "      <td>6</td>\n",
       "      <td>12</td>\n",
       "      <td>52</td>\n",
       "      <td>360</td>\n",
       "      <td>8633</td>\n",
       "      <td>517991</td>\n",
       "      <td>31079478</td>\n",
       "      <td>5.920278</td>\n",
       "      <td>120.005278</td>\n",
       "      <td>r13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2</td>\n",
       "      <td>2021-12-26 17:11:18</td>\n",
       "      <td>2021-12-26 23:06:43</td>\n",
       "      <td>2021-12-31 23:06:50</td>\n",
       "      <td>r14n27</td>\n",
       "      <td>normal</td>\n",
       "      <td>1</td>\n",
       "      <td>16</td>\n",
       "      <td>TIMEOUT</td>\n",
       "      <td>{r14n27}</td>\n",
       "      <td>...</td>\n",
       "      <td>6</td>\n",
       "      <td>12</td>\n",
       "      <td>52</td>\n",
       "      <td>360</td>\n",
       "      <td>8633</td>\n",
       "      <td>517991</td>\n",
       "      <td>31079478</td>\n",
       "      <td>5.923611</td>\n",
       "      <td>120.001944</td>\n",
       "      <td>r14</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>3</td>\n",
       "      <td>2021-12-26 17:11:18</td>\n",
       "      <td>2021-12-26 23:06:43</td>\n",
       "      <td>2021-12-31 23:06:50</td>\n",
       "      <td>r15n12</td>\n",
       "      <td>normal</td>\n",
       "      <td>1</td>\n",
       "      <td>16</td>\n",
       "      <td>TIMEOUT</td>\n",
       "      <td>{r15n12}</td>\n",
       "      <td>...</td>\n",
       "      <td>6</td>\n",
       "      <td>12</td>\n",
       "      <td>52</td>\n",
       "      <td>360</td>\n",
       "      <td>8633</td>\n",
       "      <td>517991</td>\n",
       "      <td>31079478</td>\n",
       "      <td>5.923611</td>\n",
       "      <td>120.001944</td>\n",
       "      <td>r15</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>4</td>\n",
       "      <td>2021-12-26 17:11:18</td>\n",
       "      <td>2021-12-26 23:06:43</td>\n",
       "      <td>2021-12-31 23:06:50</td>\n",
       "      <td>r10n14</td>\n",
       "      <td>normal</td>\n",
       "      <td>1</td>\n",
       "      <td>16</td>\n",
       "      <td>TIMEOUT</td>\n",
       "      <td>{r10n14}</td>\n",
       "      <td>...</td>\n",
       "      <td>6</td>\n",
       "      <td>12</td>\n",
       "      <td>52</td>\n",
       "      <td>360</td>\n",
       "      <td>8633</td>\n",
       "      <td>517991</td>\n",
       "      <td>31079478</td>\n",
       "      <td>5.923611</td>\n",
       "      <td>120.001944</td>\n",
       "      <td>r10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>5</td>\n",
       "      <td>2021-12-26 17:11:18</td>\n",
       "      <td>2021-12-26 23:06:43</td>\n",
       "      <td>2021-12-31 23:06:50</td>\n",
       "      <td>r10n30</td>\n",
       "      <td>normal</td>\n",
       "      <td>1</td>\n",
       "      <td>16</td>\n",
       "      <td>TIMEOUT</td>\n",
       "      <td>{r10n30}</td>\n",
       "      <td>...</td>\n",
       "      <td>6</td>\n",
       "      <td>12</td>\n",
       "      <td>52</td>\n",
       "      <td>360</td>\n",
       "      <td>8633</td>\n",
       "      <td>517991</td>\n",
       "      <td>31079478</td>\n",
       "      <td>5.923611</td>\n",
       "      <td>120.001944</td>\n",
       "      <td>r10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1596958</th>\n",
       "      <td>2323559</td>\n",
       "      <td>2022-11-01 13:55:57</td>\n",
       "      <td>2022-11-01 13:55:57</td>\n",
       "      <td>2022-11-01 13:58:54</td>\n",
       "      <td>r28n1</td>\n",
       "      <td>gpu_titanrtx_shared</td>\n",
       "      <td>1</td>\n",
       "      <td>6</td>\n",
       "      <td>CANCELLED</td>\n",
       "      <td>{r28n1}</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>11</td>\n",
       "      <td>44</td>\n",
       "      <td>305</td>\n",
       "      <td>7309</td>\n",
       "      <td>438595</td>\n",
       "      <td>26315757</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.049167</td>\n",
       "      <td>r28</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1596959</th>\n",
       "      <td>2323560</td>\n",
       "      <td>2022-11-01 13:56:12</td>\n",
       "      <td>2022-11-01 13:56:17</td>\n",
       "      <td>2022-11-01 13:56:24</td>\n",
       "      <td>r13n10</td>\n",
       "      <td>shared</td>\n",
       "      <td>1</td>\n",
       "      <td>6</td>\n",
       "      <td>FAILED</td>\n",
       "      <td>{r13n10}</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>11</td>\n",
       "      <td>44</td>\n",
       "      <td>305</td>\n",
       "      <td>7309</td>\n",
       "      <td>438596</td>\n",
       "      <td>26315772</td>\n",
       "      <td>0.001389</td>\n",
       "      <td>0.001944</td>\n",
       "      <td>r13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1596960</th>\n",
       "      <td>2323561</td>\n",
       "      <td>2022-11-01 13:56:45</td>\n",
       "      <td>2022-11-01 13:56:45</td>\n",
       "      <td>2022-11-01 13:58:12</td>\n",
       "      <td>r28n1</td>\n",
       "      <td>gpu_titanrtx_shared</td>\n",
       "      <td>1</td>\n",
       "      <td>6</td>\n",
       "      <td>FAILED</td>\n",
       "      <td>{r28n1}</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>11</td>\n",
       "      <td>44</td>\n",
       "      <td>305</td>\n",
       "      <td>7309</td>\n",
       "      <td>438596</td>\n",
       "      <td>26315805</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.024167</td>\n",
       "      <td>r28</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1596961</th>\n",
       "      <td>2323557</td>\n",
       "      <td>2022-11-01 13:54:21</td>\n",
       "      <td>2022-11-01 13:58:19</td>\n",
       "      <td>2022-11-01 13:59:01</td>\n",
       "      <td>r14n20</td>\n",
       "      <td>normal</td>\n",
       "      <td>1</td>\n",
       "      <td>16</td>\n",
       "      <td>COMPLETED</td>\n",
       "      <td>{r14n20}</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>11</td>\n",
       "      <td>44</td>\n",
       "      <td>305</td>\n",
       "      <td>7309</td>\n",
       "      <td>438594</td>\n",
       "      <td>26315661</td>\n",
       "      <td>0.066111</td>\n",
       "      <td>0.011667</td>\n",
       "      <td>r14</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1596962</th>\n",
       "      <td>2323558</td>\n",
       "      <td>2022-11-01 13:54:26</td>\n",
       "      <td>2022-11-01 13:58:19</td>\n",
       "      <td>2022-11-01 13:59:18</td>\n",
       "      <td>r14n21</td>\n",
       "      <td>normal</td>\n",
       "      <td>1</td>\n",
       "      <td>16</td>\n",
       "      <td>COMPLETED</td>\n",
       "      <td>{r14n21}</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>11</td>\n",
       "      <td>44</td>\n",
       "      <td>305</td>\n",
       "      <td>7309</td>\n",
       "      <td>438594</td>\n",
       "      <td>26315666</td>\n",
       "      <td>0.064722</td>\n",
       "      <td>0.016389</td>\n",
       "      <td>r14</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>1596963 rows × 23 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "              id         submit_date          start_date            end_date  \\\n",
       "0              1 2021-12-26 17:11:18 2021-12-26 23:06:31 2021-12-31 23:06:50   \n",
       "1              2 2021-12-26 17:11:18 2021-12-26 23:06:43 2021-12-31 23:06:50   \n",
       "2              3 2021-12-26 17:11:18 2021-12-26 23:06:43 2021-12-31 23:06:50   \n",
       "3              4 2021-12-26 17:11:18 2021-12-26 23:06:43 2021-12-31 23:06:50   \n",
       "4              5 2021-12-26 17:11:18 2021-12-26 23:06:43 2021-12-31 23:06:50   \n",
       "...          ...                 ...                 ...                 ...   \n",
       "1596958  2323559 2022-11-01 13:55:57 2022-11-01 13:55:57 2022-11-01 13:58:54   \n",
       "1596959  2323560 2022-11-01 13:56:12 2022-11-01 13:56:17 2022-11-01 13:56:24   \n",
       "1596960  2323561 2022-11-01 13:56:45 2022-11-01 13:56:45 2022-11-01 13:58:12   \n",
       "1596961  2323557 2022-11-01 13:54:21 2022-11-01 13:58:19 2022-11-01 13:59:01   \n",
       "1596962  2323558 2022-11-01 13:54:26 2022-11-01 13:58:19 2022-11-01 13:59:18   \n",
       "\n",
       "           node            nodetypes  numnodes  numcores      state  \\\n",
       "0         r13n5               normal         1        16    TIMEOUT   \n",
       "1        r14n27               normal         1        16    TIMEOUT   \n",
       "2        r15n12               normal         1        16    TIMEOUT   \n",
       "3        r10n14               normal         1        16    TIMEOUT   \n",
       "4        r10n30               normal         1        16    TIMEOUT   \n",
       "...         ...                  ...       ...       ...        ...   \n",
       "1596958   r28n1  gpu_titanrtx_shared         1         6  CANCELLED   \n",
       "1596959  r13n10               shared         1         6     FAILED   \n",
       "1596960   r28n1  gpu_titanrtx_shared         1         6     FAILED   \n",
       "1596961  r14n20               normal         1        16  COMPLETED   \n",
       "1596962  r14n21               normal         1        16  COMPLETED   \n",
       "\n",
       "        split_nodes  ...  submit_day_of_week month_of_year  week_of_year  \\\n",
       "0           {r13n5}  ...                   6            12            52   \n",
       "1          {r14n27}  ...                   6            12            52   \n",
       "2          {r15n12}  ...                   6            12            52   \n",
       "3          {r10n14}  ...                   6            12            52   \n",
       "4          {r10n30}  ...                   6            12            52   \n",
       "...             ...  ...                 ...           ...           ...   \n",
       "1596958     {r28n1}  ...                   1            11            44   \n",
       "1596959    {r13n10}  ...                   1            11            44   \n",
       "1596960     {r28n1}  ...                   1            11            44   \n",
       "1596961    {r14n20}  ...                   1            11            44   \n",
       "1596962    {r14n21}  ...                   1            11            44   \n",
       "\n",
       "         day_of_year  hour_of_year minute_of_year  second_of_year  \\\n",
       "0                360          8633         517991        31079478   \n",
       "1                360          8633         517991        31079478   \n",
       "2                360          8633         517991        31079478   \n",
       "3                360          8633         517991        31079478   \n",
       "4                360          8633         517991        31079478   \n",
       "...              ...           ...            ...             ...   \n",
       "1596958          305          7309         438595        26315757   \n",
       "1596959          305          7309         438596        26315772   \n",
       "1596960          305          7309         438596        26315805   \n",
       "1596961          305          7309         438594        26315661   \n",
       "1596962          305          7309         438594        26315666   \n",
       "\n",
       "         job_waiting_time  job_duration  rack  \n",
       "0                5.920278    120.005278   r13  \n",
       "1                5.923611    120.001944   r14  \n",
       "2                5.923611    120.001944   r15  \n",
       "3                5.923611    120.001944   r10  \n",
       "4                5.923611    120.001944   r10  \n",
       "...                   ...           ...   ...  \n",
       "1596958          0.000000      0.049167   r28  \n",
       "1596959          0.001389      0.001944   r13  \n",
       "1596960          0.000000      0.024167   r28  \n",
       "1596961          0.066111      0.011667   r14  \n",
       "1596962          0.064722      0.016389   r14  \n",
       "\n",
       "[1596963 rows x 23 columns]"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "mark_ml_df(df)\n",
    "\n",
    "# process column 'state'\n",
    "df['state'] = df['state'].apply(lambda x: 'CANCELLED' if x is not None and x.startswith('CANCELLED') == True else x)\n",
    "\n",
    "df[\"submit_day\"] = df[\"submit_date\"].dt.date\n",
    "df[\"submit_hour_of_day\"] = df[\"submit_date\"].dt.hour\n",
    "df[\"submit_day_of_week\"] = df[\"submit_date\"].apply(lambda x:x.weekday())  # dt.weekday\n",
    "\n",
    "df['month_of_year'] = df[\"submit_date\"].dt.month\n",
    "df['week_of_year'] = df[\"submit_date\"].apply(lambda x: x.strftime(\"%U\"))\n",
    "df['day_of_year'] = df[\"submit_date\"].apply(lambda x: x.timetuple().tm_yday)\n",
    "df[\"hour_of_year\"] = df[\"submit_date\"].apply(lambda x: int((x - datetime(x.year, 1, 1)).total_seconds() / 3600))\n",
    "df[\"minute_of_year\"] = df[\"submit_date\"].apply(lambda x: int((x - datetime(x.year, 1, 1)).total_seconds() / 60))\n",
    "df[\"second_of_year\"] = df[\"submit_date\"].apply(lambda x: int((x - datetime(x.year, 1, 1)).total_seconds()))\n",
    "\n",
    "df[\"job_waiting_time\"] = df[\"start_date\"] - df[\"submit_date\"]\n",
    "df[\"job_waiting_time\"] = df[\"job_waiting_time\"].apply(lambda x:timedelta.total_seconds(x))\n",
    "df[\"job_waiting_time\"] = df[\"job_waiting_time\"]/60/60 # hour\n",
    "\n",
    "df[\"job_duration\"] = df[\"end_date\"] - df[\"start_date\"]\n",
    "df[\"job_duration\"] = df[\"job_duration\"].apply(lambda x:timedelta.total_seconds(x))\n",
    "df[\"job_duration\"] = df[\"job_duration\"]/60/60 # hour\n",
    "\n",
    "df[\"rack\"] = df[\"node\"].str.slice(0,3)\n",
    "\n",
    "generic_df = df.loc[df[\"is_ml\"] == 0]\n",
    "ml_df = df.loc[df[\"is_ml\"] == 1]\n",
    "df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "ge_completed = generic_df[generic_df['state'] == 'COMPLETED']\n",
    "ge_failed = generic_df[generic_df['state'] == 'FAILED']\n",
    "ml_completed = ml_df[ml_df['state'] == 'COMPLETED']\n",
    "ml_failed = ml_df[ml_df['state'] == 'FAILED']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 2400x500 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# fig-generic-ml-failures-day-of-hour\n",
    "from matplotlib.pyplot import MultipleLocator\n",
    "\n",
    "fig, axes = plt.subplots(figsize = (24, 5), nrows = 1, ncols = 2)\n",
    "fig.gca().yaxis.set_major_formatter(matplotlib.ticker.StrMethodFormatter('{x:,.0f}'))\n",
    "\n",
    "x_vals = np.arange(ge_failed[\"submit_hour_of_day\"].nunique())\n",
    "y_vals = ge_failed[\"submit_hour_of_day\"].value_counts(sort=False)\n",
    "axes[0].bar(x_vals, y_vals, edgecolor='black', capsize=5, width = 0.6, color = 'steelblue')\n",
    "axes[0].set_ylim(bottom=0, top=None)\n",
    "axes[0].set_xlim(-0.5,23.5)\n",
    "axes[0].set_xlabel(\"Hour of Day\",  fontsize=28)\n",
    "axes[0].set_ylabel(\"Number of Failed Jobs\", fontsize=28)\n",
    "axes[0].tick_params(axis='both', which='major', labelsize=26)\n",
    "axes[0].tick_params(axis='both', which='minor', labelsize=26)\n",
    "axes[0].xaxis.set_major_locator(MultipleLocator(1))\n",
    "\n",
    "\n",
    "x_vals = np.arange(ml_failed[\"submit_hour_of_day\"].nunique())\n",
    "y_vals = ml_failed[\"submit_hour_of_day\"].value_counts(sort=False)\n",
    "axes[1].bar(x_vals, y_vals, edgecolor='black', capsize=5, width = 0.6, color = 'lightcoral')\n",
    "axes[1].set_ylim(bottom=0, top=None)\n",
    "axes[1].set_xlim(-0.5,23.5)\n",
    "axes[1].set_xlabel(\"Hour of Day\", fontsize=28)\n",
    "axes[1].set_ylabel(\"Number of Failed Jobs\", fontsize=28)\n",
    "axes[1].tick_params(axis='both', which='major', labelsize=26)\n",
    "axes[1].tick_params(axis='both', which='minor', labelsize=26)\n",
    "axes[1].xaxis.set_major_locator(MultipleLocator(1))\n",
    "\n",
    "\n",
    "#fig.gca().yaxis.set_major_formatter(matplotlib.ticker.StrMethodFormatter('{x:,.0f}'))\n",
    "plt.show()\n",
    "# fig.savefig('fig-generic-failures-week-and-day.pdf', bbox_inches = 'tight')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 900x300 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# fig-generic-ml-failures-day-of-hour\n",
    "from matplotlib.pyplot import MultipleLocator\n",
    "\n",
    "fig, axes = plt.subplots(figsize = (9, 3))\n",
    "fig.gca().yaxis.set_major_formatter(matplotlib.ticker.StrMethodFormatter('{x:,.0f}'))\n",
    "\n",
    "x_vals = np.arange(ge_failed[\"submit_hour_of_day\"].nunique())\n",
    "y_vals = ge_failed[\"submit_hour_of_day\"].value_counts(sort=False)\n",
    "axes.bar(x_vals, y_vals, edgecolor='black', capsize=5, width = 0.6, color = 'steelblue')\n",
    "axes.set_ylim(bottom=0, top=None)\n",
    "axes.set_xlim(-0.5,23.5)\n",
    "axes.set_xlabel(\"Hour of Day\",  fontsize=20)\n",
    "axes.set_ylabel(\"Number of Failed Jobs\", fontsize=20)\n",
    "axes.tick_params(axis='both', which='major', labelsize=18)\n",
    "axes.tick_params(axis='both', which='minor', labelsize=18)\n",
    "axes.set_xticks(np.arange(1, 22, step=2))\n",
    "axes.xaxis.set_major_locator(MultipleLocator(2))\n",
    "\n",
    "\n",
    "plt.show()\n",
    "fig.savefig('plots/section_5/fig-generic-failures-hour-of-day.pdf', bbox_inches = 'tight')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 900x300 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# fig-generic-ml-failures-day-of-hour\n",
    "from matplotlib.pyplot import MultipleLocator\n",
    "\n",
    "fig, axes = plt.subplots(figsize = (9, 3))\n",
    "fig.gca().yaxis.set_major_formatter(matplotlib.ticker.StrMethodFormatter('{x:,.0f}'))\n",
    "\n",
    "x_vals = np.arange(ml_failed[\"submit_hour_of_day\"].nunique())\n",
    "y_vals = ml_failed[\"submit_hour_of_day\"].value_counts(sort=False)\n",
    "axes.bar(x_vals, y_vals, edgecolor='black', capsize=5, width = 0.6, color = 'lightcoral')\n",
    "axes.set_ylim(bottom=0, top=None)\n",
    "axes.set_xlim(-0.5,23.5)\n",
    "axes.set_xlabel(\"Hour of Day\",  fontsize=20)\n",
    "axes.set_ylabel(\"Number of Failed Jobs\", fontsize=20)\n",
    "axes.tick_params(axis='both', which='major', labelsize=18)\n",
    "axes.tick_params(axis='both', which='minor', labelsize=18)\n",
    "axes.set_xticks(np.arange(1, 22, step=2))\n",
    "axes.xaxis.set_major_locator(MultipleLocator(2))\n",
    "\n",
    "\n",
    "plt.show()\n",
    "fig.savefig('plots/section_5/fig-ml-failures-hour-of-day.pdf', bbox_inches = 'tight')"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "base",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.12.5"
  },
  "orig_nbformat": 4
 },
 "nbformat": 4,
 "nbformat_minor": 2
}